The overall objective of this study is to demonstrate that treatment with the drug granulocyte- macrophage colony-stimulating factor (GM-CSF) can reduce the incidence of nosocomial infection in high-risk, critically injured children. Traumatic injury remains, by far, the leading cause of death for children outside the neonatal period in the United States. The incidence of nosocomial infection is extremely high in injured children who require ICU care and it represents an important source of morbidity and health care costs. Impairment of innate immune function is common following critical injury in adults and is characterized by a reduced capacity of whole blood to produce the pro-inflammatory cytokine tumor necrosis factor (TNF)-1 upon ex vivo stimulation with bacterial lipopolysaccharide (LPS) and reduced monocyte HLA-DR expression. We have developed the capacity to perform highly standardized, generalizable, same-day functional immune monitoring in our laboratory and our preliminary data show that children with an ex vivo LPS-induced TNF1 production capacity < 600 pg/ml are at particularly high risk for the development of nosocomial infection. Our pediatric preliminary data, along with several small adult studies, suggest that GM-CSF can reverse critical illness-induced immunodepression. GM-CSF is FDA-approved for use in children and has a low side-effect profile, but has never been studied in critically injured children. We therefore propose a series of trials including a prospective, single-center, randomized, double-blind, placebo-controlled trial of GM-CSF in critically injured children screened for severe innate immune suppression. The studies outlined in this submission are designed to test the novel central hypothesis that immunomodulation with GM-CSF will result in reduction in the risk of nosocomial infection after critical injury in high-risk children through safe, rapid, and sustained improvement in innate immune function. The GM-CSF for Immunomodulation Following Trauma (GIFT) study will be the largest study to investigate the in vivo use of an innate immunostimulant for the treatment of critical illness-induced immune suppression in any patient population. For all Aims of this project we will screen severely injured children (those with an Injury Severity Score > 10) for innate immune suppression as defined by an ex vivo LPS-induced TNF1 production capacity < 600 pg/ml. Only those subjects with severe innate immune suppression will be further evaluated in the GIFT study. For our first Specific Aim we will perform a series of dose-escalation studies in cohorts determined by age and presence or absence of severe traumatic brain injury (TBI) in order to determine the lowest immunostimulatory, tolerable dose (LITD) of GM-CSF that will safely improve immune function (ex vivo LPS-induced TNF1 production capacity and monocyte HLA-DR expression) to levels that were not associated with nosocomial infection in our preliminary studies. We will go on to use these LITDs in Specific Aim 2 in which we will perform a prospective, single-center, randomized, double-blind, placebo-controlled trial of GM-CSF for the prevention of nosocomial infection in children without severe TBI. We will also perform a single-arm trial for the smaller population of children with severe TBI, also with the outcome measure of nosocomial infection risk. In Specific Aim 3 we will evaluate the relationships between five measures of innate immune function (cytokine production capacity, antigen presenting capacity, migration, phagocytosis, and microbial killing) with infection risk, with GM-CSF responsiveness, and with each other. We anticipate that the GIFT study will represent a paradigm shift in the management of pediatric trauma in that it will demonstrate the role of immune stimulation in reducing infection after pediatric critical injury, will show the feasibility of real-time immune function monitoring, and will yield the largest and most comprehensive set of immune function data of any trauma population yet studied.